Mobile Network Access for MVNOs Market Assessment

Version 1.0

30 January 2015

Prepared for:

Cartesian: Mobile Network Access for MVNOs

Version History

Version Date Author Changes Distribution Approval 1.0 30/1/2015 Cartesian Final version incorporating UCC feedback UCC T Twinemanzi

Confidentiality

This document and the information contained herein are private and confidential, and are solely for the use of the Uganda Communications Commission.

Copyright

The contents of this document are copyright © 2015 Cartesian Ltd. All rights reserved.

The information contained herein is the property of Cartesian and is provided on condition that it will not be reproduced, copied, lent, or disclosed, directly or indirectly, nor used for any purpose other than that for which it was specifically furnished.

Cartesian Ltd. Registered in England and Wales. Registered Number: 03230513 Registered Office Address: Descartes House, 8 Gate Street, London WC2A 3HP United Kingdom

Copyright © 2014 Cartesian Ltd. All rights reserved. 1

Cartesian: Mobile Network Access for MVNOs

Table of Contents

1. Executive Summary ...... 4 Conclusions of SMP assessment ...... 5 2. Introduction ...... 5 3. Relevant market definition and description ...... 6 4. Market Performance Assessment ...... 7 Demand for MNA ...... 7 Penetration of MVNO services in the retail market ...... 9 Potential for Market Growth in the retail market ...... 9 Prices (Cost of Access) ...... 10 Service Quality ...... 11 Innovation ...... 11 5. Market Structure Assessment ...... 12 Market Concentration ...... 12 Network Effects and Externalities ...... 13 Barriers to entry ...... 14 Sunk Costs ...... 14 Economies of Scale ...... 14 Economies of Scope ...... 15 Extent of Vertical Integration ...... 15 6. Market Conduct Assessment ...... 15 Scale and Ability to Access Resources ...... 15 Control of Essential Upstream Inputs ...... 15 Access to Sales and Distribution Channels ...... 16 Transparency ...... 16 Ease of Consumer Switching ...... 17 Countervailing Buyer Power ...... 17 Evidence of Dynamic Competition ...... 17 Joint Dominance ...... 18 7. Basic Market Conditions ...... 18 Technology ...... 18 Cost Conditions ...... 20 8. Overall Conclusion of SMP assessment ...... 21

Copyright © 2014 Cartesian Ltd. All rights reserved. 2

Cartesian: Mobile Network Access for MVNOs

List of Figures

Figure 1: Number of Subscriptions and Market Share in Uganda by MNO, (2014 Q3) ...... 8 Figure 2: Mobile Revenue Forecast in Uganda, 2011 – 2019...... 9 Figure 3: Mobile Subscriptions Forecast in Uganda, 2011 – 2019 ...... 10 Figure 4: Price Terms for K2-Orange MVNO Agreement, 2014 ...... 11 Figure 5: Mobile Subscriptions Market Share in Uganda, 2011 – 2014 ...... 12 Figure 6: Market Concentration (HHI) in Uganda Retail Mobile Services, 2011 – 2014 ...... 13 Figure 7: Simplified Architecture of Full MVNO ...... 19 Figure 8: Simplified Architecture of Light MVNO ...... 20

Copyright © 2014 Cartesian Ltd. All rights reserved. 3

Cartesian: Mobile Network Access for MVNOs

1. Executive Summary

This document sets out an assessment of whether or not there is effective competition in the market for mobile network access for MVNOs, (the “MNA” market), particularly whether or not one or more of the mobile network operators (MNOs) have significant market power (SMP) in the supply of mobile network access to mobile virtual network operators (MVNOs). A firm or group of MNOs with SMP have limit entry or raise the cost of entry of MVNOs into the markets of retail mobile services, which would either increase prices above and/or reduce output below the competitive level. It follows that consumers will benefit less in a market where MNOs exercise SMP compared with a MNA market in which competition is effective.

MNA refers to the provision of access to a mobile operator’s network and systems for the purpose of hosting an MVNO. The services in this market are supplied on a national basis to licensed mobile CSPs without mobile network and systems infrastructure (i.e. MVNOs).

Currently there is only one MVNO in Uganda, which is , and it has an MNA agreement with the Orange. In addition to Orange, other MNOs operating in Uganda include MTN, Airtel, , and .

In our view, the other MNOs have less economic incentive to supply MNA to an MVNO than Orange. This assessment is based on the following factors:

 Orange has a relatively small number retail customers, which means that an MVNO provides Orange with an opportunity to grow the total number of customers (Orange retail customers + K2 retail customers) carried on its network;  There are limited risks of cannibalisation as Orange has a small retail market share;  We believe Orange has spare network capacity which it can make available to an MVNO; and consequently,  Hosting an MVNO can be expected to deliver a net increase in revenues and profit.

In contrast, the fact that MTN and Airtel have a relatively large customer base implies that they are at greater risk of losing these customers to an MVNO. Furthermore, they may have less spare network capacity as a result of carrying more customers.

Smart Telecom is a new entrant with a small customer base, but it is more likely to target the same customer base that a MVNO new entrant would, and its network may not be sufficiently developed to carry an MVNO.

Finally, Uganda Telecom's network is likely to be less attractive to an MVNO than that of the other MNOs. This is primarily due to its lack of a national data network, which would inhibit the creation of competitive retail offerings.

With respect to the demand of MNA by MVNOs, there appears to be little or no demand beyond K2, which has already entered the market.

In summary, the current scope of the MNA market consists of one MVNO, which is K2, and one MNO, which is Orange. It should be noted, however, this is a "narrow" definition of the market. It could be argued that the scope of the market should also include MTN and Airtel, because, even though they currently do you supply MNA services, they could do so, and might have the incentive to do so in the long run, should their circumstances change. That is, a

Copyright © 2014 Cartesian Ltd. All rights reserved. 4

Cartesian: Mobile Network Access for MVNOs

"broader" definition of the market would include MTN, Airtel and Orange on the supply side, with K2 on the demand side.

Conclusions of SMP assessment Our initial finding with respect to the narrow market definition is that Orange does not have significant market power in the MNA market. This assessment is based on the observation that Orange and K2 both freely entered into the MNA. In other words, both K2 and Orange were able to negotiate contract terms that were sufficiently beneficial to them so that they could reach an agreement. In particular, it appears that K2 had sufficient countervailing buyer power to offset Orange's market power. This is supported by the observation that Orange could have foreclosed K2's market entry, but allowed K2 enter into the market, which suggests that Orange expects to benefit from K2's entry.

If the market definition were expanded, so that K2 had the option to purchase MNA from Orange, Airtel, or MTN, then this would reduce further the risk of any of these operators having SMP. This follows from the preceding assessment of the narrow market definition, as the inclusion of Airtel and MTN it introduces additional competitive constraint, which would reduce the MNO's market power.

2. Introduction

The purpose of this document is to set out an assessment of whether or not there is effective competition in market for mobile network access for MVNO, (the “MNA” market), particularly whether or not one or more of the mobile network operators (MNOs) have significant market power (SMP) in the supply of MNA to mobile virtual network operators (MVNOs).

MNOs with SMP have the potential to limit entry of MVNOs in the market. This implies that customers in downstream retail markets would not derive the same levels of benefits from this market as they would if there were effective competition in the upstream market for MNA. Therefore, it is important to identify whether or not firms have SMP.

If one or more MNOs are found to have SMP, then it may be appropriate to introduce regulatory measures. Properly designed and implemented, regulatory measures can ameliorate SMP and thus enhance market performance and economic efficiency in the relevant market.

This assessment commences with a description of the definition of the MNA market. This is followed by an assessment of the market performance and pricing structure and levels, amongst other factors. Market structure is then examined, which addresses recent developments and the concentration of market shares amongst MNOs and MVNOs. This is followed by an assessment of the conduct of MNOs and MVNOs in the market, and in turn a description of the basic market conditions. This report concludes with our initial overall conclusions regarding this assessment of SMP in the market for mobile network access for MVNO within Uganda.

Copyright © 2014 Cartesian Ltd. All rights reserved. 5

Cartesian: Mobile Network Access for MVNOs

3. Relevant market definition and description

The relevant market for CSPI is defined in the report prepared by Cartesian, for the Uganda Communications Commission (UCC), Review of Uganda Communications Markets: Relevant Markets Report, dated 29th October 2014 (the “Relevant Markets Report”). The Relevant Markets Report identifies that the MNA market as a priority for the SMP assessment.

The key features of the MNA market identified in the Relevant Markets Report are:

 The MNA market is a wholesale market;  MVNOs purchase access to an MNO's mobile network so that they can supply mobile services to retail customers; and  The geographic scope of the market encompasses the whole of the geographic region of Uganda and does not differentiate between regions within Uganda.

Currently there is only one MVNO in Uganda, which is K2 Telecom, and it has an MNA agreement with the Orange. In addition to Orange, other MNOs operating in Uganda include MTN, Airtel, Uganda Telecom, and Sure Telecom.

Overview of relevant market

A generally accepted definition of an MVNO is that of a licenced CSP that provides mobile communications services, without holding a spectrum licence of its own and without necessarily holding all the core network elements and/or infrastructure required to provide these mobile communications services.

As MVNOs do not have direct access to spectrum, they must negotiate an agreement with an MNO that provides the technical and commercial conditions for access to a host MNO’s network.

The MNA market is still nascent in Uganda since the first and only MVNO, K2 Telecom, was launched in 2013. K2 Telecom offers consumers in the retail market a set of mobile telephony and data services (such as voice calls, SMS and internet access) by using Orange Uganda’s network infrastructure.

MVNOs thus operate in the same downstream retail market as the MNO they have a wholesale MNA agreement with, as well as the other MNOs operating in the retail market. The extent to which an MVNO competes with its host MNO will governed by the terms of the wholesale arrangement between the MVNO and the MNO. These terms might include the processes and procedures for implementing the MVNO's retail prices.

In the case of the MNA agreement between K2 and Orange, as this agreement was voluntary – i.e. Orange was not compelled to provide K2 with access to its network – then, we expect that Orange Uganda engaged with K2 to enhance rather erode its total value. For instance, the MNA agreement may enhance Oranges total value if Orange had under-utilised network capacity and it believed that K2 could help grow its total revenues (= retail + wholesale revenue) faster than it could on its own.

Copyright © 2014 Cartesian Ltd. All rights reserved. 6

Cartesian: Mobile Network Access for MVNOs

We expect, though, that an MVNO would endeavour to compete against those MNOs that it does not have a MNA agreement with. Again, with respect to the K2-Orange agreement, Orange may believe that K2 would assist to compete with the other MNOs.

In general, we expect MVNOs to differentiate their services from their host MNO in aspects that they retain control of – such as price, distribution, VAS, and branding – that allow the MVNO to target customer segments that the host MNO might not target. However, the MVNO cannot differentiate its service from the host MNO through differences in network coverage or data speeds, or anything that is entirely dependent on the quality of the host MNO’s network.

The coverage, performance and functionality of an MNO’s network are therefore important considerations for prospective MVNOs in selecting their network partner. Other factors to consider include the service wrap, compatibility regarding operating model, the ability to avoid channel conflict (e.g. targeting the same customers as the host MNO), contract terms and price.

While there is only one example of MNA agreement in the Ugandan market, in the following discussion we use to explore not only issues specific to the K2-Orange MNA agreement, but also to identify any issues with the functioning of the MNA market more generally.

4. Market Performance Assessment

Demand for MNA It is necessary to understand the economic incentives of MNOs and MVNOs for engaging in an MNA agreement in order to determine the demand for MNA agreements.

As noted above, there is currently one MVNO in Uganda, which is K2, and K2 has an MNA agreement with Orange. In addition to Orange, there are four other MNOs in Uganda: MTN, Airtel, Uganda Telecom, and Sure Telecom. This suggests, on the surface, that there are in total 5 MNOs that might supply MNA. Of these 5 MNOs, the limited extent of Uganda Telecom’s W-CDMA would not be attractive to an MVNO that wanted to bundle voice, SMS and mobile data. This would reduce the number of potential host MNOs to Orange, MTN, Airtel, and Sure Telecom.

Assuming that there are 4 potential host MNOs, then this might suggest that there are a large number of options for an MVNO, such as K2, to consider. However, whether or not an MNO chooses to host an MVNO, will depend on the MNO's expectation of costs and benefits from MNA an arrangement. Likewise, an MVNO will consider its costs and benefits from entering into the retail mobile market in Uganda.

A key criteria that an MVNO is likely to consider if it is assessing entry into the retail market is whether it is profitable. An MVNO is unlikely to enter the retail market if it does not expect to earn a positive return. A factor that will influence whether or not it expects to earn a positive return will be the number of communications services providers (CPSs) in the retail market. The fact that there are currently 6 CPSs in total – i.e. 5 MNOs and 1 MVNO –in the retail market suggests that an MVNO entrant is likely to encounter potentially significant rivalry, which increases the risk that it would not earn a positive return. This in turn would reduce its incentive to enter into, and thus demand for, an MNA agreement.

Copyright © 2014 Cartesian Ltd. All rights reserved. 7

Cartesian: Mobile Network Access for MVNOs

Likewise, we expect that an MNO's incentive to engage with an MVNO would be governed by its expectation of whether the MVNO would increase its total profitability. That is, the MNO is likely to consider whether the MNVO would increase its revenues by more than its costs.

With respect to an MNO's revenue, the MNO is likely to consider whether the MVNO will increase its wholesale revenue without reducing its retail revenue. This would occur if the MVNO builds its own customer base without reducing the host MNO's customer base. Whether or not an MNO is likely to consider that an MVNO will increase or decrease its revenue is likely to depend on the size of its current customer base. If the MNO has a large (small) customer base then an MVNO is more (less) likely to reduce its customers and thus revenues.

With respect to an MNO’s costs, hosting an MVNO places increased demands on network capacity. In evaluating the business case for hosting an MVNO, an MNO will therefore seek to ensure that its network costs are more than covered by the price it charges for wholesale access. An MNO with an underutilised network may choose to offer lower access charges than one that is capacity constrained.

In summary, if the MNO has an underutilised network and relatively small retail market share, it is likely to be more predisposed to hosting an MVNO. In relation to the MNA market in Uganda, Orange fits these criteria.

The following table summarises the number of mobile subscriptions, as measured by the number of active SIMs, and the market shares for each of the CSPs. It highlights that Airtel and MTN have a significantly larger customer base than the other for CSPs. Given this information, the preceding discussion would imply that Airtel and MTN are to have less incentive to engage with a MVNO than the three smaller MNOs. That is, an MNVO is likely to pose a greater threat to their customer base.

Figure 1: Number of Subscriptions and Market Share in Uganda by MNO, (2014 Q3) Mobile Voice and SMS Mobile Operator Market Share Subscriptions MTN 10,050,782 51.3% Airtel 7,562,729 38.6% Uganda Telecom 1,016,385 5.2% Orange 683,153 3.5% Sure Telecom 124,881 0.6%

Source: UCC

Furthermore, of the three smaller MNOs, the limited extent of Uganda Telecom's W-CDMA network, means that it would be less attractive to an MVNO that wanted to bundle voice, SMS and mobile data. Sure Telecom is a new entrant that is currently establishing itself. This means that it is unlikely to have the network capacity in the near to medium term to support an

Copyright © 2014 Cartesian Ltd. All rights reserved. 8

Cartesian: Mobile Network Access for MVNOs

MVNO, and also there is a possibility that it will be targeting the same customer segments that a MVNO entrant might target.

Therefore, this leaves Orange Uganda as the most likely host MNO. As the above table indicates, the size of its current customer base is small relative to MTN and Airtel. Also, it was revealed during interviews that it has recently invested in and expanded its network capacity. Taken together, these factors suggest that Orange had sound economic incentives to supply MNA to K2.

The presence of a single MVNO in the Ugandan mobile market raises the question whether there is demand for MNA that is not currently being met. During interviews with the CSPs, none indicated that there were other potential MVNOs seeking MNA. This may indicate that demand is currently being met by the market. One possible explanation is that prospective MVNOs do not find the retail mobile market in Uganda to be sufficiently attractive for market entry.

Penetration of MVNO services in the retail market By August 2014, K2 Telecom had around 220,000 of the approximately 19 million subscriptions in Uganda, which represents close to 1% of the mobile market (by subscriptions).

Potential for Market Growth in the retail market Mobile voice services are forecast to generate approximately USD 689M in revenue for CSPs by the end of 2014. This is a reduction in revenue when compared to the USD 684M in 2012 (see Figure 2). Voice revenues are forecast to decrease significantly over the next year and will continue to fall to reach USD 593M by 2019.

Figure 2: Mobile Revenue Forecast in Uganda, 2011 – 2019

1,200

1,000

502 800 300 404 166 224 76 94 122

600

U$ U$ Millions 400 684 686 689 684 674 654 616 593 200

- 2012 2013 2014 2015 2016 2017 2018 2019

Mobile Voice Revenues Mobile Data Revenues

Source: Ovum, 2014

Copyright © 2014 Cartesian Ltd. All rights reserved. 9

Cartesian: Mobile Network Access for MVNOs

In terms of subscriptions (i.e. active SIMs), the mobile market has been growing at a CAGR (2011-2014) of 10% There are currently more than 19 million mobile subscriptions in the country, which is equivalent to 50% of the population. Continued growth of subscriptions is predicted at annual rate of approximate 7% over the next 5 years (CAGR, 2014-2019). It is estimated that by 2019, there will be approximately 28 million mobile subscriptions by the end of 2019, representing a significant growth from the current 19 million subscriptions (Figure 3).

Figure 3: Mobile Subscriptions Forecast in Uganda, 2011 – 2019

Actual Forecast 30 28 27 25 25 23 21 19 20 20 18 17 16

15 Millions 10

5

-

Source: Cartesian, Ovum and UCC

However, this observed growth in subscriptions is likely to be caused by users purchasing additional SIM cards rather than growth in the actual number of users of mobile services. The fact that mobile voice and SMS services have peaked and are now starting to decline, would support the proposition that the growth in subscriptions has largely been driven by users purchasing additional SIM cards.

Uganda still has significant potential for retail mobile service penetration so more MVNOs with disruptive propositions are possible.

Furthermore, operators have plans to invest in coverage, capacity and new technologies (e.g. LTE). This will improve the options available to MVNOs as a result.

Prices (Cost of Access) MVNOs that purchase MNA from an MNO must pay for set-up costs required to facilitate the hosting arrangement. The MVNO must then pay for wholesale access on an ongoing basis. As there is only one example of an MVNO in Uganda, we will consider the price terms agreed between K2 and Orange in this section.

Copyright © 2014 Cartesian Ltd. All rights reserved. 10

Cartesian: Mobile Network Access for MVNOs

K2 pays Orange UGX 60 per voice call minute. The fee for per SMS is UGX 15, while data is charged at UGX 0.3 per kilobyte. These fees have not been renegotiated since the initial contract was agreed.

In addition to the UGX 60 per call minute, Orange charges K2 a transit fee of UGX 25 per minute for calls carried to other networks, on top of any termination fee set for K2 by MNOs. This transit fee substantial reduces K2’s voice call margins.

These terms are summarised in Figure 4 below.

Figure 4: Price Terms for K2-Orange MVNO Agreement, 2014 Cost Item Fee per Unit (UGX) Voice minute 60 per minute Voice transit 25 per minute SMS 15 per SMS Data 0.3 per kilobyte

Source: Stakeholder Data

We understand that Orange will often charge K2 for the full wholesale value of retail bundles sold to its customers regardless of the actual usage. For example, a customer buying 100 on- net minutes for UGX 30,000 and using 50 minutes in the designated time period would result in a cost charged to K2 for UGX 6,000 (100 x UGX 60), even though half of the minutes were not used.

As discussed in the section on transparency, price and product information for MNA services are not available publicly. Additionally, rate cards are neither requested by nor shared with the UCC.

After the takeover of the Orange Uganda brand by Africell, K2 are in the process of renegotiating the terms of their MVNO agreement to achieve what they view as a fair revenue share. They are particularly keen to reduce the cost of transit services provided by Orange.

Service Quality Orange and K2 have agreed SLAs in the MNA contract that specify requirements for service quality measures such as network uptime, call quality and request response time. However, interviews with K2 have indicated that these are not enforced and it is believed that Orange could do more to improve their network and service quality.

Innovation The issues noted above regarding the quality of the MNA services suggest that innovations can be made to improve service quality. Furthermore, as will be discussed below in the section on basic market conditions, the MNA can take a range of forms from a "full MVNO" to a "light MVNO", which suggests further opportunities for innovation in the MNA market.

Copyright © 2014 Cartesian Ltd. All rights reserved. 11

Cartesian: Mobile Network Access for MVNOs

5. Market Structure Assessment

Market Concentration The suppliers to the market are the licenced MNOs that have spectrum licences and access to mobile network infrastructure. These suppliers also offer retail mobile services. In Uganda, these are MTN Uganda, , Orange Uganda, Uganda Telecom, and Sure Telecom.

MTN Uganda is part of the international MTN group, with operations in 22 countries.

Another significant CSP is Airtel Uganda, which is part of Bharti Airtel group with operations in Africa and Asia. Airtel acquired Warid in 2013, becoming the second largest CSP in Uganda.

In 2014, Orange Uganda was acquired by Africell, an African group with presence also in the Gambia and Sierra Leone. Previously, Orange Uganda had been part of the international Orange S.A. group (formerly known as France Telecom).

Sure Telecom Uganda Limited entered the market in early 2014 under the brand name “Smart”. Sure Telecom is owned by the Aga Khan Fund for Economic Development.

MTN is the market leader in the downstream retail mobile market with 51% market share by subscriptions as of September 2014 (Figure 16). The second largest CSP, Airtel, has 39% of subscriptions. The remaining 8% of the market is shared among UTL (less than 5%), Orange (3%), K2 (less than 1%) and Sure Telecom (less than 1%).

Figure 5: Mobile Subscriptions Market Share in Uganda, 2011 – 2014

MTN Airtel Warid UTL K2 Sure Orange

60%

51% 51% 49% 50% 50% 48% 49% 48% 48% 50% 46% 46% 47% 47% 46% 47% 41% 39% 39% 40%

30% 26% 27% 27% 25% 26% 26% 25% 25% 23% 24% 21% 22% 19% 19% 20% 17% 16% 17% 16% 15% 15% 16% 16% 16% 16%

8% 8% 7% 10% 7% 6% 6% 6% 5% 5% 5% 5% 4% 5% 5% 5% 7% 5% 6% 6% 7% 6% 6% 0% 5% 5%

Source: UCC

Copyright © 2014 Cartesian Ltd. All rights reserved. 12

Cartesian: Mobile Network Access for MVNOs

As can be seen from Figure 5, the retail market is highly concentrated. In particular, Airtel and MTN have significant market shares, of 39% and 51% respectively. Hence, 90% of the retail mobile call and SMS market is held between these two operators.

Given the distribution of market share, the level of market concentration, as measured by the Herfindahl-Hirschman Index, is unsurprisingly very high (Figure 16). In 2011, the HHI value was already in excess of 3,000. This increased through 2012 and 2013 due to share gains by MTN and Airtel. In 2013, there was a sharp increase due primarily to the Airtel/Warid merger and, to a lesser extent, the increase in Airtel and MTN’s market shares at the expense of the smaller operators. In the 9 months since 2013, there has been a small decrease in HHI due to the fall in Airtel’s share.

Figure 6: Market Concentration (HHI) in Uganda Retail Mobile Services, 2011 – 2014

5,000

4,500

4,000

3,500

3,000

2,500

Source: UCC, Cartesian

Based on current trends and market conditions, the concentration level is unlikely to change significantly within the timeframe of this assessment (3 to 5 years). MTN and Airtel are likely to keep the largest share of the retail market since it is challenging for either a small player or a new entrant to challenge MTN and Airtel’s position.

Network Effects and Externalities There are no direct network effects or externalities in the MNA market. The value of a host MNO’s network to an MVNO is not determined by the number of other MVNOs that are hosted on it.

Copyright © 2014 Cartesian Ltd. All rights reserved. 13

Cartesian: Mobile Network Access for MVNOs

Barriers to entry The main barrier to entry for suppliers in the MNA market is the requirement to have an operational mobile network. Establishing and operating a mobile network requires allocation of suitable spectrum and access to significant funding.

For existing mobile CSPs, entry into the MNA market then requires additional investment in systems and processes to host MVNOs.

The technology required to add MVNO-hosting functionality to an existing network is commercially available from several suppliers. Compared to the costs of establishing the mobile network, these development costs are relatively small. There is no need for the MNO to secure additional spectrum or communication licenses.

We therefore conclude that there are no material barriers for existing mobile CSPs to enter the MNA market.

In the downstream retail market, the key barrier to entry for an MVNO is network access. As the UCC does not explicitly mandate non-discriminatory access to MVNOs, there have been cases where some CSPs have taken an extended period of time to negotiate access or even denying access altogether to fair and reasonable requests. We consider these issues below under “control of essential inputs”.

Sunk Costs There are significant sunk costs in establishing a mobile network, for example in network electronics, IT systems and other infrastructure. These costs are incurred by the CSP primarily to enter the retail mobile markets for voice, SMS and data. Whilst this expenditure also enables the provision of MNA, this is secondary to the primary aim of retail market entry.

With regard to the specific provision of MNA, the sunk costs associated with hosting MVNOs are small in comparison with the overall network costs.

Mobile CSPs can thus incur a small cost to increase the profitability of other existing assets by offering MNA. As such, MNOs are incentivised to supply access seekers with network services to grow their revenues.

This criteria is thus relevant to the assessment of SMP in this market, as CSPs can benefit from increasing the utilisation of their existing assets through offering MNA.

Economies of Scale MNOs invest large sums in order to build and operate mobile networks for the primary purpose of providing retail telephony and data services to customers.

Through data obtained from K2 and the UCC, we estimate that K2’s total voice traffic is approximately 3 to 4% of Orange’s total voice traffic at present, a small fraction of network demand.

Having invested in building the network, the incremental capacity costs of additional voice and data traffic is marginal. Thus, offering MNA benefits from the scale economics inherent in mobile networks.

Copyright © 2014 Cartesian Ltd. All rights reserved. 14

Cartesian: Mobile Network Access for MVNOs

Economies of Scope MNA is an incremental service that leverages existing infrastructure used by the MNO in other retail services that it offers consumers. These are the mobile network (radio and core elements) and spectrum. The mobile network is used in providing retail telephony and data services to consumers.

There are thus very few costs involved in adding MNA functionality to the system and these may already be in place for use by the MNO itself if it has an in-house MVNO.1

Thus, the MNO’s operation of network elements to offer other retail services means that offering MNA is incremental to their existing service portfolio and leverages common assets. It would thus only be offered by an MNO that is also offering retail telephony and data services to consumers.

Extent of Vertical Integration The potential suppliers of MNA are also suppliers in the downstream retail mobile market. In the retail market, the CSPs compete with the MVNOs that rely on them for network access.

As MVNOs usually only realistically intend to obtain a market share of less than 10% (based on international benchmarks), this competition is usually acceptable to both MVNOs and access providers so long as the terms are fair and reasonable to both parties.

Also, many MVNOs choose to focus on serving a particular market segment which may not be well served by the host MNO. Indeed, MNOs may seek to host MVNOs that are complementary as a means of cost-effectively reaching underserved segments.

Nevertheless, the vertical integration creates a potential conflict of interest between the wholesale network services business (offering MNA) and the retail business of the MNO.

We therefore find this criterion relevant to the market assessment.

6. Market Conduct Assessment

Scale and Ability to Access Resources The resources required to provide the MNA are readily available to all MNOs and as such, we find this criteria to not be relevant for this SMP analysis.

Control of Essential Upstream Inputs MNA is an upstream input to the retail mobile voice and data services markets in Uganda. As discussed in the section on Vertical Integration, all potential suppliers of MNA also supply this downstream retail market, i.e. MNOs that supply MNA also compete with MVNOs in the downstream market.

1 There are several international examples of MNO’s launching services under a secondary (or sub-) brand. Although these secondary brands are in-house from the perspective of the MNO, it is common practice for them to be operated as MVNOs.

Copyright © 2014 Cartesian Ltd. All rights reserved. 15

Cartesian: Mobile Network Access for MVNOs

There is therefore a potential conflict of interest. MNOs may seek to deny MNA to MVNOs or otherwise restrict the use of MNA as a means of avoiding additional competition in the downstream retail market.

Collectively, the licenced MNOs control access to the MNA market. If all of the MNOs refused to supply MNA, then it would not be possible for an MVNO to enter the downstream retail market. However, an MVNO only requires one MNO to provide MNA for it to enter the retail market and K2 was able to secure an MNA agreement with Orange.

Beyond simply denying MNA to prospective MVNOs, MNOs may impose conditions on MVNOs that hamper their ability to compete in the retail market.

In Uganda, K2 is a “light MVNO”. This means that instead of investing in its own service platform, it relies on use of Orange’s platform. As a result, this places constraints on K2’s ability to set its own tariffs. These must be sent in advance to Orange, and discussions with K2 have indicated a significant lead time (several weeks) in order to change tariffs. In the prepaid market, this is a significant challenge for K2 as they are unable to react to market dynamics. Orange has some incentive to ensure that K2 cannot compete with them directly, and so may delay changes to ensure that K2 is less reactive than them in the retail market.

Additionally, MNOs will process call detail records of customers on behalf of light MVNOs (such as K2). This opens up the possibility for them to identify high-value customers and poach them by advertising their own services directly to the customer.

As such, this is a relevant and applicable criteria for this SMP analysis.

Access to Sales and Distribution Channels From a supplier perspective, MNA suppliers are known to all potential MVNOs.

As such, we find this criteria is not relevant to the SMP analysis.

Transparency Transparency in the MNA market is poor: none of the MNOs publish reference offers for MNA supply; the terms and conditions of MNA supply contracts are not disclosed; and, in Light MVNO agreements, the MVNOs are dependent upon the MNOs for calculation and reconciliation of usage fees.

From a contract perspective, the terms and conditions of MNA agreements are not made public. Indeed, rate cards and traffic data are not submitted to the UCC by Orange for the MNA services offered to K2.

As the UCC intervened in the negotiations between Orange and K2, the UCC has some visibility into the terms negotiated between the two companies. However, there is limited visibility of changes to tariffs made by K2 and the performance of Orange as a host MNO in providing MNA services to K2 after this initial negotiation was completed.

A second transparency issue relates to the reconciliation of call data and volumes. For Light MVNOs this is driven primarily by the MNOs as the MNOs hold the billing records. As such, MVNOs may have limited visibility of access provider data in order to verify any claims made by the host MNO.

Copyright © 2014 Cartesian Ltd. All rights reserved. 16

Cartesian: Mobile Network Access for MVNOs

As MNA service costs are determined on the basis of traffic generated, there is a risk that an MNA providers may seek to overstate the amount of traffic generated by MVNOs. This would then allow the supplier to overcharge the MVNO.

For example, K2 have indicated that they are charged the full amount for bundles sold to but not used by the customer. For a 100 voice minute bundle, K2 are charged for the 100 minutes sold rather than the actual amount used by the customer (which may be less than 100 minutes).

As such, we find transparency to be relevant and applicable in the analysis of SMP in this market.

Ease of Consumer Switching The switching costs for an MVNO to change host MNO are significant. Consequently, there are very few international examples of MVNOs switching supplier.

For Light MVNOs, such as K2, the reliance upon the host MNOs operational and billing systems makes switching complex and costly. Furthermore it requires cooperation and support from the incumbent supplier which may be difficult to obtain.

One of the greatest impediments to switching supplier is the need to reissue SIM cards to retail customers. While this may not be costly in material terms, the effort required and potential for customer churn in the process is highly significant.

As such, MVNOs may find themselves locked in to an access provider. This gives the MNO the ability to squeeze an MVNO’s margins over time.

We thus find this criteria to be relevant and applicable to the SMP analysis.

Countervailing Buyer Power As noted, at the present time there is one MNA agreement between Orange and K2, where another MNO is unlikely to have the economic incentive to enter into a MNA agreement with K2, and there does not appear to be any other firms seeking to enter into the retail market as an MVNO.

In these circumstances, K2's negotiating power with Orange is largely determined by its countervailing buyer power. In other words, understanding the benefits that K2 can bring to Orange as a result in improvements in the price and non-price terms of the MPA agreement.

The fact that K2 and Orange freely entered into a MNA agreement suggests that both CSPs expect to benefit from it. This also indicates that K2 had sufficient countervailing buyer power to negotiate terms that were acceptable. One possible way that K2 might improve on these terms would be to develop a better understanding of how Orange benefits from the MPA, in order to grow its countervailing buyer power, and use this information to gain better terms for itself.

Evidence of Dynamic Competition As discussed above, K2 faces high barriers to switching and the other MNOs in the market appear less amenable or suited to hosting an MVNO. We therefore do not expect that other MNOs will attempt to compete with Orange to supply K2.

Copyright © 2014 Cartesian Ltd. All rights reserved. 17

Cartesian: Mobile Network Access for MVNOs

Also, K2 is the only MVNO in Uganda and there is no indication that many more MVNOs are likely in the near term. There is therefore limited opportunity for suppliers in the market to engage in competition.

These factors suggest that there is little scope for dynamic competition in the MNA market to emerge in the near term.

Joint Dominance We found no evidence to indicate that MNOs exercise joint dominance in wholesale market for MNA. In particular, the fact that MTN and Airtel have significant market share in the retail market does not mean that they dominant in the market for MNA. The fact that they have not engaged in the MNA market is likely to be the result of a lack of economic incentive to enter into the market and the small likelihood of further MVNOs entering into the market.

7. Basic Market Conditions

Technology Suppliers offering MNA are MNOs, i.e. mobile CSPs with spectrum licences and their own mobile network infrastructure. The technologies offered in MNA are the same as those required to provide retail mobile services to end-users.

The technology required to provide (and acquire) MNA is well established, and will not change significantly in the foreseeable future.

There are several operational types within the MVNO category, based on the split of network elements and roles.

These components are as follows:

 Mobile Switching Centre (MSC) and Gateway Mobile Switching Centre (G-MSC): MSCs are telephone switches that make the connection between mobile users within the network (MSC), from mobile users to the public switched telephone network (MSC), and from mobile users to other mobile networks (G-MSC)  Gateway GPRS support node (GGSN): Responsible for the internetworking between the GPRS networks and external packet switched networks, such as the Internet. PDN Gateway (P-GW) may also be used for the same purpose as GGSN in LTE networks.  Serving GPRS support node (GGSN): The SGSN is the service access point to the GPRS network for the mobile user. The SGSN performs the same functions as the MSC.  Service Delivery Platform (SDP): Platform allows CSPs to deliver application-based and cross-technology services, including SMS, MMS, USSD, IVR, and content VAS.  Short Message Service – Service Centre (SMSC): Responsible for handling the SMS operations of a wireless network. MMSC is the equivalent for MMS.  Customer Relationship Management (CRM): Integrated management information system that is used to schedule, plan and control the sales and pre-sales activities in an organisation. It includes billing and fulfilment systems used to process call detail records into airtime balance changes in real-time and to support the provisioning of customer SIM cards.

Copyright © 2014 Cartesian Ltd. All rights reserved. 18

Cartesian: Mobile Network Access for MVNOs

 Home Location Register (HLR): Central database that contains details of each mobile phone subscriber authorised to use the network.  Visitor Location Register (VLR): Temporary database associated with an MSC that contains details of each mobile phone subscriber authorised to use the network. Database entry of the subscriber is deleted when the subscriber leaves the service area.

A full MVNO holds control, to the greatest extent, over the services it provides in the retail market and operates its own core network elements used for the service provision (e.g. HLR, VLR, MSC, EIR, AuC, SMSC, SGSN) This allows for higher flexibility in designing and implementing new services. A full MVNO can issue its own SIM cards, but it does not hold usage rights over the radio frequencies and, therefore, nor does it hold elements of the radio access network (e.g. base stations, radio network controllers).

Figure 7: Simplified Architecture of Full MVNO

Source: Cartesian

A light MVNO does not hold any core network infrastructure, but it controls the relationship with the customer. Most MVNOs in this category are simply distributors. Within this operating model, an MVNO can issue its own SIM cards using its own brand name.

Copyright © 2014 Cartesian Ltd. All rights reserved. 19

Cartesian: Mobile Network Access for MVNOs

Figure 8: Simplified Architecture of Light MVNO

Source: Cartesian

There are other types of MVNOs that lie between the extremes of the light and full types. They differ based on the services offered and the degree of independence from the mobile network operator. In international practice, this “medium” MVNO category is further split in two subcategories:

 Service Provider MVNOs: Compared to light MVNOs, these MVNOs also manage all the processes regarding the customer relations services and the billing services for the services provided;  Value-Added Service (VAS) MVNOs: Compared to service provider MVNOs, VAS MVNOs hold certain infrastructure elements, which ensure control over the services provided. These MVNOs can easily provide value-added services such as voice messages, missed call notifications, virtual private networks (VPN) etc.

K2 operates as a light MVNO in Uganda, and while it retains control over its customer relationships (including provisioning and billing), it does not have direct control over network routing, tariffs or the provision of VAS.

Cost Conditions The incremental cost of supplying MNA is relatively small compared to the investment that an MNO must make to build out its network for retail mobile market. In fact, it is these high

Copyright © 2014 Cartesian Ltd. All rights reserved. 20

Cartesian: Mobile Network Access for MVNOs

costs that drive the decision for an MVNO to seek MNA rather than incur significant fixed costs to acquire spectrum and build their own network.

In Uganda, spectrum fees are not high by international standards. Spectrum fees are fixed, based on specific frequency ranges, and determined on a price per MHz basis. For example, 10 MHz on the 1,900 MHz band was awarded for a one-time fee of UGX 100,000,000.

However, the cost of network build the network is considerable. MTN Uganda’s CAPEX in 2013 was approximately USD 50.8M (compared to USD 40M in 2012), which represents 12% of its revenue in the country. Most of this expenditure was to improve GSM and W-CDMA network coverage and quality. This exemplifies the magnitude of the investment required.

MNOs also incur some fixed costs to host MVNOs, though this is relatively small when compared to their overall costs, and is largely borne by the MVNO.

By comparison the fixed costs for MVNOs are relatively small, including for example, software and systems to operate a retail mobile service for their customers. The majority of costs for an MVNO are the variable costs it must pay its host MNO for network access.

Finally, MVNOs will need to acquire all their customers through advertising and other means, which may also be significant for a new entrant if it has no existing customer base to leverage.

In some cases, an existing brand operating in a non-telecoms retail market (e.g. consumer goods retailer, lifestyle brand, postal services) may launch an MVNO and leverage their existing customer base and/or distribution channels to acquire customers and obtain market visibility and share.

8. Overall Conclusion of SMP assessment

Currently there is only one MVNO in Uganda, which is K2 Telecom, and it has an MNA agreement with the Orange. This suggests that the current scope of the MNA market consists of one MVNO, which is K2, and one MNO, which is Orange. It should be noted, however, this is a "narrow" definition of the market. It could be argued that the scope of the market should also include MTN and Airtel, because, even though they currently do you supply MNA services, they could do so, and might have the incentive to do so if in the long run their circumstances change. That is, a "broader" definition of the market would include MTN, Airtel and Orange on the supply side, with K2 on the demand side.

Our initial finding with respect to the narrow market definition is that Orange does not have significant market power in the MNA market. This assessment is based on the observation that Orange and K2 both freely entered into the MNA. In other words, both K2 and Orange were able to negotiate contract terms that were sufficiently beneficial to them so that they could reach an agreement. In particular, it appears that K2 had sufficient countervailing buyer power to offset Orange's market power. This is supported by the observation that Orange could have foreclosed K2's market entry, but allowed K2 enter into the market, which suggests that Orange expects to benefit from K2's entry.

If the market definition were expanded, so that K2 had the option to purchase MNA from Orange, Airtel, or MTN, then this would reduce further the risk of any of these operators having SMP. This follows from the preceding assessment of the narrow market definition, as

Copyright © 2014 Cartesian Ltd. All rights reserved. 21

Cartesian: Mobile Network Access for MVNOs

the inclusion of Airtel and MTN it introduces additional competitive constraint, which would reduce the MNO's market power.

In addition to Orange, other MNOs operating in Uganda include MTN, Airtel, Uganda Telecom, and Sure Telecom. In our view, the other MNOs have less economic incentive to supply MNA to an MVNO than Orange. This assessment is based on the following factors:

 Orange has a relatively small number retail customers, which means that an MVNO provides Orange with an opportunity to grow the total number of customers (Orange retail customers + K2 retail customers) carried on its network;  There are limited risks of cannibalisation as Orange has a small retail market share;  We believe Orange has spare network capacity which it can make available to an MVNO; and consequently,  Hosting an MVNO can be expected to deliver a net increase in revenues and profit.

In contrast, the fact that MTN and Airtel have a relatively large customer base implies that they are a greater risk of losing these customers to an MVNO. Furthermore, they may have less spare network capacity as a result of carrying more customers.

Sure Telecom is a new entrant with a small customer base, but it is more likely to target the same customer base that a MVNO new entrant would, and its network may not be sufficiently developed to carry an MVNO.

Finally, Uganda Telecom's network is likely to be less attractive to an MVNO than that of the other MNOs. This is primarily due to its lack of a national data network, which would inhibit the creation of competitive retail offerings.

With respect to the demand of MNA by MVNOs, there appears to be little or no demand beyond K2, which has already entered the market.

Copyright © 2014 Cartesian Ltd. All rights reserved. 22

Cartesian is a specialist consulting firm of industry experts, focused exclusively on the communications, technology and digital media sector. For over 20 years, Cartesian has advised clients in strategy development and assisted them in execution against their goals. Our unique portfolio of professional services and managed solutions are tailored to the specific challenges faced by executives in these fast- moving industries. Combining strategic thinking and practical experience, we deliver superior results.

www.cartesian.com For further information, please contact us at [email protected]